static¶
- torch_sim.runners.static(system, model, *, trajectory_reporter=None, autobatcher=False)[source]¶
Run single point calculations on a batch of systems.
Unlike the other runners, this function does not return a state. Instead, it returns a list of dictionaries, one for each batch in the input state. Each dictionary contains the properties calculated for that batch. It will also modify the state in place with the “energy”, “forces”, and “stress” properties if they are present in the model output.
- Parameters:
system (StateLike) – Input system to calculate properties for
model (ModelInterface) – Neural network model module
unit_system (UnitSystem) – Unit system for energy and forces
trajectory_reporter (TrajectoryReporter | dict | None) – Optional reporter for tracking trajectory. If a dict, will be passed to the TrajectoryReporter constructor and must include at least the “filenames” key. Any prop calculators will be executed and the results will be returned in a list. Make sure that if multiple unique states are used, that the variable_atomic_numbers and variable_masses are set to True in the state_kwargs argument.
autobatcher (BinningAutoBatcher | bool) – Optional autobatcher to use for batching calculations
- Returns:
Maps of property names to tensors for all batches
- Return type: